EP1271357A2 - Verfahren und Programm zur Verwaltung von Profilen zum Filtern von Informationen - Google Patents

Verfahren und Programm zur Verwaltung von Profilen zum Filtern von Informationen Download PDF

Info

Publication number
EP1271357A2
EP1271357A2 EP02250562A EP02250562A EP1271357A2 EP 1271357 A2 EP1271357 A2 EP 1271357A2 EP 02250562 A EP02250562 A EP 02250562A EP 02250562 A EP02250562 A EP 02250562A EP 1271357 A2 EP1271357 A2 EP 1271357A2
Authority
EP
European Patent Office
Prior art keywords
profile
profiles
user
information
information filtering
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Withdrawn
Application number
EP02250562A
Other languages
English (en)
French (fr)
Other versions
EP1271357A3 (de
Inventor
Y. c/o Hitachi Ltd Intel. Property Group Inaba
K. c/o Hitachi Ltd Intellectual Prop. Gp Tada
Y. c/o Hitachi Ltd Intellectual Prop. Gp Sato
T. c/o Hitachi Ltd Intel. Prop. Gp Matsubayashi
M. c/o Hitachi Ltd Intel. Prop. Group Uchikado
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hitachi Ltd
Original Assignee
Hitachi Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hitachi Ltd filed Critical Hitachi Ltd
Publication of EP1271357A2 publication Critical patent/EP1271357A2/de
Publication of EP1271357A3 publication Critical patent/EP1271357A3/de
Withdrawn legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation

Definitions

  • the present invention relates to a profile management method for information filtering and more specifically to a profile management method for obtaining the desired result of users without any duplication and omission on the occasion of filtering document information using a profile as the search condition.
  • a large amount electronic documents (hereinafter, referred to as texts) are distributed to users from time to time in the form of electronic mail (E-mail) and electronic news or the like in recent years.
  • E-mail electronic mail
  • an information source for originating information pieces by utilizing WWW World Wide Web
  • the amount of texts collected from these information sources using information collecting robots or the like is also increasing intensively. Therefore, recently, there is intensive rise of needs for information filtering system to search the texts including the information pieces which users want and to distribute such information pieces obtained to users.
  • Such information filtering system is disclosed, for example, in the official gazette of the Japanese Unexamined Patent Publication No. HEI 10-27182 (hereinafter, referred to as the cited reference 1).
  • This cited reference 1 refers to the technique for distributing only the document information conforming to the query expression formed of words preset by a user to the relevant user.
  • search noise the search result includes many documents not conforming the object of search.
  • the applicant of the present invention have proposed the Japanese Patent Application No. HEI 1-75005 (hereinafter referred to as the cited reference 2) as a filtering technique to improve this problem explained above.
  • the user inputs, in place of the words, a sample document (hereinafter referred to as a "seed document") indicating the information who wants distribution thereof as the search condition.
  • seed document a sample document
  • the conformity of contents between such seed document and document information for distribution is calculated with the predetermined method and only the document information pieces having conformity exceeding the predetermined value are distributed to the relevant user.
  • Fig. 2 shows a PAD (Problem Analysis Diagram) indicating the processes of the filtering system utilizing an ordinary profile.
  • Fig. 3 is a schematic diagram showing the practical flows of processes of the filtering system utilizing an ordinary profile.
  • a search condition data (hereinafter referred to as "profile") of the relevant user is generated (S502).
  • profile search condition data
  • document information is transmitted from an information resource (S503), the conformity between the profile of each user and the relevant document information is calculated and the relevant document information is distributed to the user satisfying the predetermined condition (S504).
  • a user 201 registers a sample document 202 (hereafter after referred to as a "seed document") indicating the information to be distributed.
  • seed document a sample document 202
  • the user 201 desires distribution of the information in regard to High school baseball championship when it is generated and therefore this user sets the seed document 202 "High school baseball championship has been opened at the Koushien ball-park etc.
  • the system retrieves a keyword (hereinafter referred to as "characteristic string”) which indicates, as the characteristic, the contents of such document from the seed document 202, counts up the number of times of appearing of such keywords in the seed document and registers the counted data as a pair to a profile 203 as the weight of each characteristic string (1 ⁇ in Fig. 3).
  • characteristic string such as "High School”, “Baseball”, “Koushien” and “Opened” are retrieved from the seed document 202 and the number of times of appearing of such characteristic strings are registered to the profile 203 as the weights.
  • S(D) is conformity between document information D and profile
  • Frq (i) is the number of times of appearing of characteristic string i in the document information D
  • w(i) is a weight of the characteristic string i within the relevant profile.
  • ⁇ indicates a sum of all characteristic strings in the relevant profile.
  • the cited reference 2 relates to a technique of information filtering for previously generating a profile based on the seed document with which the user is capable of searching the object document from a large amount of documents.
  • Fig. 4 is a conceptual diagram showing flows of practical processes in the case where the user has a plurality of profiles in the filtering system using ordinary profiles.
  • the user In general, the user often has the interest in various objects and therefore it is desirable for the user that various topics of a plurality of objects can be distributed. Therefore, it is general to introduce the system that the user is capable of setting a plurality of profiles and the information filtering system distributes, to the relevant user, only the document information pieces conforming to respective profiles.
  • the user 308 wants the document information about professional baseball and the document information about high school baseball and therefore sets respectively the profile A304 and profile B305.
  • the professional baseball and high school baseball are different objects, but these are common topics in the field of baseball and therefore the profile A304 and the profile B305 are similar in the contents.
  • the characteristic strings “baseball” and “opened” are registered to both profiles A and B. Accordingly, the document information including these characteristic strings has higher conformity with both profiles A and B.
  • the duplicated document information pieces are included in the document information 306 conforming to the profile A and in the document information 307 conforming to the profile B.
  • the document information pieces "High school baseball championship has opened at the Koushien ball-park" and "Japan series of professional baseball has opened at......"include the character strings of "baseball” and "opened"
  • the conformity between the profiles A and B is high and therefore these document information pieces are distributed in duplication.
  • the user 308 desires distribution of topic of Olympic games at Sydney and therefore sets the profile C309.
  • the topic of the Olympic games at Sydney has becomes the topic in the past and any topic about this Olympic games is no longer generated.
  • any means is not provided for the user to detect by himself/herself that the user is still setting an old and useless profile.
  • the user can set many profiles, it is very difficult for the user to detect each time whether the user is still setting useless profiles or not.
  • the number of profiles which may be set by the user is limited and in such a service system in which charging is executed based on the number of profiles being set, it is a serious problem for the user that there is no means to detect existence of such useless profiles.
  • the present invention has been proposed to address the problems explained above and it is therefore preferable that the present invention provides a profile management method which can obtain excellent distribution result without any noise and omission for information filtering by detecting existence of a plurality of similar profiles and existence of old profiles and then informing it to the user in the information filtering for presenting the document information through the filtering using the profile as the search condition data.
  • the present invention to provide a profile management method which can prevent holding of useless profiles by adequately and easily optimizing and deleting the useless profiles such as old profiles, permit the user to effectively set the profiles and moreover does not result in deterioration of performance as the system without referring to useless profiles.
  • the invention in relation to the profile management method used for the information filtering of the present invention is structured to calculate validity for the information filtering to the profile and to notify the validity of profile to the user depending on the calculation result of validity in the profile management method to be used for the information filtering for presenting the document information through the filtering by calculating the conformity with the profile as the search condition data.
  • validity of the information filtering may be defined as similarity among profiles in the case where a plurality of profiles are designated.
  • two or more profiles designed with the user are integrated into a profile among a plurality of profiles explained above.
  • the invention in relation to the profile management method used for the information filtering of the present invention may also be structured additionally to instruct the user to designate two or more profiles by calculating the conformity with the profile as the search condition data in the profile management method used for the information filtering presented by filtering of document information, to compare characteristics of respective profiles designated and correct the profiles through specialization to provide a difference in the result at the time of information filtering.
  • validity explained above is calculated, in the profile management method used for information filtering, on the basis of the generation frequency of the text information matched with a profile of the information in the case where the information filtering is executed in the past within the predetermined period, user evaluation of the text information matched with a profile of the information in the case where the information filtering is executed in the past within the predetermined period, correction frequency of profile executed by the user in the past within the predetermined period and generation date of the relevant profile.
  • Fig. 1 is a system configuration diagram of the information filtering system of the present invention.
  • the information filtering system of the present invention comprises a display 100, a keyboard 101, a central processing unit (CPU) 102, a main memory 104 and a bus 103 connecting these elements.
  • CPU central processing unit
  • the bus 103 is also extended to a document information distribution source 106 for distributing document information via a communication line 105 such as LAN (Local Area Network to which a user 107 utilizing the information filtering system is connected) or the like.
  • the document information distribution source 106 distributes electronic document information to this system using an electronic mail (E-mail) and presents document information via the Internet.
  • E-mail electronic mail
  • the user 107 registers the search condition to this system using E-mail.
  • This system distributes the document information searched based on the search condition to the relevant user using an E-mail.
  • This preferred embodiment will be explained below under the assumption that the document information distribution source 106 distributes document information to this system using an E-mail or the like but it is also possible that the document information distribution source 106 is given only the function to present information on the Internet and therefore collection of texts is executed using an information collection robot.
  • the user 107 registers the search condition to this system using an E-mail but it is also possible that the user registers such search condition to this system via the Web.
  • the text explained above searched based on the search condition is distributed to the relevant user from this system using an E-mail, but it is also possible to down-load such texts via the Internet or the like.
  • the main memory 104 stores, for execution, a main control program 110, a search profile generation program 120, a document information distribution control program 121, a profile monitor program 122, a profile integration program 123, a program specialization program 124, a profile deletion program 125, a user profile storing area 129, an inter-profile similarity table 130 and a weekly hit table of each profile 131.
  • the profile monitor program 122 of these programs is formed of an inter-profile similarity monitor program 126 and a profile validity monitor program 127.
  • Fig. 5 is a PAD (Problem Analysis Diagram) showing the process sequence of the main control program 110.
  • the main control program 110 is driven when an instruction is received from a keyboard 101 of a system administrator of the information filtering system.
  • the main control program 110 drives, when it is determined that a seed document is inputted from the user 107 (S401), a search profile generation program 120 to generate a search profile of the relevant user (S402).
  • a search profile generation program 120 drives, when it is determined that a seed document is inputted from the user 107 (S401), a search profile generation program 120 to generate a search profile of the relevant user (S402).
  • the practical method for generating a search profile is same as that explained in the paragraph the related art.
  • the main control program 110 drives the document information distribution control program 121, calculates the conformity between the profile of each user and the relevant document information and then distributes the relevant document information to the user satisfying the predetermined condition (S404).
  • a practical calculation method of conformity is also identical to that described in the paragraph of the related art.
  • the profile monitor program 122 is driven to calculate the inter-profile similarity and profile application period which are set with the user and an alarm is then notified to the user 107 (S406).
  • the practical calculation method will be explained later.
  • the profile integration program 123 is driven to generate a profile having integrated contents of a plurality of profile designated with the relevant user and then delete the original profile (S408).
  • the practical profile integration sequence will be explained later.
  • the profile specialization program 124 is driven to respectively correct the contents of a plurality of profiles designated with the relevant user to the specialized contents (S410).
  • the practical profile specialization method will be explained later.
  • a profile deletion program 125 is driven to delete the profile designated with the relevant user from the user profile storing area 129 (S412).
  • the practical profile deletion method will be explained later.
  • the process 1 ⁇ of Fig. 3 is indicated as the typical process.
  • the search profile generation method it is assumed that the characteristic string is retrieved from the seed document 202, the number of times of appearing of such characteristic string is counted and such count value is written into the profile 203 as a weight of the characteristic string, as explained in the paragraph of the related art.
  • the characteristic string may be retrieved with the method explained in the paragraph of the related art or with the morphological analysis using a word dictionary.
  • a weight the number of times of appearing of each character string in the seed document 202 is defined but it is also possible to define the other indices.
  • the process 2 ⁇ of Fig. 3 is indicated as the typical process.
  • the document information of various contents is distributed from the information resource 204. Thereafter, conformity between such document information and profile of each user stored in the user profile storing area is calculated.
  • conformity S (D) is calculated using the equation (1).
  • Fig. 6 is a PAD showing the process sequence of the profile monitor program 122.
  • This program refers contents of profiles of the user and notifies similarity of contents to the user having a plurality of similar profiles in view of urging such user to optimize the profiles. Moreover, this program searches also whether the user has the profiles including old contents and not indicating the recent interest of the people or not and issues a warning to the users having such profiles.
  • the profile monitor program 122 drives the inter-profile similarity monitor program 126, searches whether each user has set or not a plurality of similar profiles and then issues a warning to the user having set such profiles (S701).
  • the profile monitor program 122 drives the profile validity monitor program 127, searches whether each user has set or not the invalid profiles such as those indicating the old topics and then issues a warning to the user having set such profiles (S702).
  • the user can attain the distribution result without any noise and omission, because existence of a plurality of similar profiles and old profiles can be detected and this fact can also be notified to the user by calling, with the profile monitor program 122 as explained above, the inter-profile similarity monitor program 126 and profile validity monitor program 127.
  • Fig. 7 is a PAD showing the process sequence of the inter-profile similarity monitor program 126.
  • Fig. 8 is a schematic diagram showing the flows of practical processes of the inter-profile similarity monitor program 126.
  • This inter-profile similarity monitor program 126 is driven with the profile monitor program 122 to determine whether there are similar profiles or not among the profiles being set with the user.
  • This program moreover calculates similarity indicating in what degree the profiles are similar and issues a warning to the user when there are similar profiles.
  • similarity is an index indicating in what degree profiles are similar among a plurality of profiles and can be thought as a validity of a plurality of profiles. Namely, when similarity is large, validity of a plurality of profiles can be evaluated as small and when similarity is small, validity of a plurality of profiles can be evaluated as large.
  • the inter-profile similarity monitor program 126 is executed repetitively for all users to which the processes of the steps S802 to S804 are registered (S701).
  • step S803 As the first process, the process of step S803 is repeated for all profiles being set with the user (S802).
  • the process of step S803 calculates, for all profiles, the similarity with all of the other profiles being set with the relevant user with the predetermined method.
  • a certain user 107 sets three profiles of the profile A, profile B and profile C to the user profile storing area 129.
  • the profile similarity monitor program 126 calculates similarity among these three profiles with the (equation 2) and writes such similarity into an inter-profile similarity table 131 (S901).
  • the profile A and profile B include the common characteristic strings such as "baseball” and "opening", the higher similarity can be obtained from the (equation 2).
  • the profiles of which similarity exceeds the predetermined value are retrieved from the inter-profile similarity table 131 (S902).
  • the profile A and profile B are retrieved because the similarity between these profiles exceeds the predetermined value (for example, a degree of similarity is 50).
  • a set of the profile A and profile B retrieved in the step S902 is presented to the user 107 together with the information that "these profiles are similar"(S903).
  • a comment "the profile A and profile B indicated below are similar” is displayed on the display screen 904 together with contents of respective profiles.
  • the user 107 knows the fact that the user has set two profiles having the similar contents.
  • Fig. 9 is a PAD showing the process sequence of the profile validity monitor program 127.
  • Fig. 10 is a schematic diagram showing flows of practical processes of the profile validity monitor program 127.
  • This profile validity monitor program 127 is driven in the step S702 with the profile monitor program 122 to determine whether there exists useless profiles for the user for which any information hitting to such profiles because these are already old is no longer generated or not among the profiles being set with each user and to notify the fact to the user as a warning when there exists such useless profiles.
  • the profile validity monitor program 127 repeats the processes of the steps S1002, S1003 for all users being registered (S1001).
  • the validity of each profile being set with the relevant user is calculated with the predetermined method.
  • the validity means an index to indicate in what effectiveness the filtering of document information can be executed. For example, when the document information corresponding to a certain profile is often generated recently, the validity of this profile is set to a higher value under the condition that the topic of such profile is rather new. However, on the contrary, when the document information corresponding to such profile is not generated recently, the validity of this profile is set to a lower value by assuming that such profile has a higher possibility as an "old and useless" profile.
  • a certain user 107 has already set three profiles of the profile A, profile B and profile C which are stored in the user profile storing area 129. Moreover, a value of the number of hits of each profile counted whenever the document information matched with each profile is generated within the last one week is held in the weekly hit table 131 of each profile.
  • the profile validity monitor program 127 refers first to the weekly hit table of each profile to retrieve the profiles of which number of hits is lower than the predetermined value (S1101).
  • the profile retrieved in the step S1101 is presented to the user 107 as the profile having a higher validity (S1102).
  • the number of hits of the profile C in the last one week is "0" and this value is lower than the predetermined value. Therefore, a warning "Your profile C indicated below is already old, isn't it?" is displayed on the display screen 1103 together with the contents of the profile C.
  • the user 107 can detect existence of the profiles which are already old and cannot generate any related topics.
  • the information of the last one week is stored in the weekly hit table 131 of each profile but it is also possible to store the information of the other period in place of one week depending on the type of application of the system.
  • a means which is different from that explained above may also be used as a means for calculating the validity.
  • Fig. 11 is a PAD showing the process sequence of the profile integration program 123.
  • Fig. 12 is a schematic diagram showing the flows of practical processes of the profile integration program 123.
  • This profile integration program 123 is driven in the step S408 with the main control program 110 when the user inputs the instruction to "integrate" the similar profile presented by the inter-profile similarity monitor program 126.
  • This profile integration program 123 can be effectively used to eliminate useless profiles by integrating a plurality of similar profiles into one profile to prevent duplicated distribution to the user 107 of the document information which is matched with respective profiles.
  • the profile integration program 123 first reads contents of a plurality of profiles designated from the user 107 (S1201).
  • a profile generated in the step S1202 is set as a profile of the relevant user and the profile read in the step S1201 is deleted (S1203).
  • the profile integration program 123 first reads contents of the profile A1310 and the profile B1311 designated by the user 107 (S1301).
  • this program 123 generates a new profile D1312 in which the weights of the characteristic strings in each profile are added (S1302).
  • the weights of the respective characteristic strings being set in common to the profile A and profile B are added and the added weight is written into a new profile D with the weight of the characteristic strings included in only one profile left as it is.
  • the integration means as explained above is used but the other means may also be used.
  • the weights of the characteristic strings which are set in common in both profiles A and B that only larger weight is set in direct as the weight of the new profile D.
  • an average value of the weights of the characteristic strings being set in the profile A and profile B may be set.
  • Fig. 13 is a PAD showing the process sequence of the profile specialization program 124.
  • Fig. 14 is a schematic diagram showing the flows of the practical processes of the profile specialization program 124.
  • This profile specialization program 124 is driven when the user 107 inputs an instruction to "specialize” the similar profiles presented with the inter-profile similarity monitor program 126 to "the profiles specialized to respective contents" .
  • This profile specialization program 124 can effectively be used to prevent duplicated distribution to the user 107 of document information pieces matched with respective profiles by specializing a plurality of similar profiles to respective contents and thereby to distribute, instead, the desired document information to the user 107 without any omission.
  • the profile specialization program 124 first reads contents of a plurality of profiles designated with the user 107 (S1401).
  • contents of a plurality of profiles read in the step S1401 are corrected to the contents specialized in respective profiles with the predetermined method (S1402).
  • S1402 An example of the process to correct respective profiles to the specialized profiles will be explained later.
  • the user 107 is given a warning with the inter-profile similarity monitor program 126 that the preset profile A1510 and the profile B1511 are similar profiles.
  • the user 107 inputs a request to correct the similar profiles to those specialized to respective contents, upon determination that if these similar profiles are left as they are, a fear for duplicated distribution of the document information pieces matched with respective profiles is generated.
  • the profile specialization program 124 first reads contents of the designated profile A1510 and profile B1511 (S1501).
  • a negative weight is given, in the step S1502, to the characteristic string which is included in the profile B1511 but in the profile A1510 and this weight is added to the profile A1510.
  • the characteristic strings "High school” and "Koushien” included in the profile B1511 are added to the profile A1510 by giving a negative weight thereto.
  • a negative weight is given in the step S1503 to the characteristic string which is included in the profile A1510 but in the profile B1511 and this weight is added to the profile B1511.
  • the characteristic strings "professional” and “league” included in the profile A1510 are added to the profile B1511 by giving a negative weight.
  • the methods other than that explained above may be used.
  • the profile specialization program 124 is capable of preventing the distribution of duplicated document information pieces by respectively correcting contents of a plurality of similar profiles to the specialized contents and then using such profiles for the information filtering.
  • Fig. 15 is a PAD showing the process sequence of the profile deletion program 125.
  • This profile deletion program 125 is driven when the user 107 inputs an instruction for "deletion" of the profiles which are presented and determined to be invalid with the profile validity monitor program 127.
  • This program can be used effectively to prevent, when a topic becomes sufficiently old and document information in regard to such topic is no longer generated, that the old and useless profiles are still maintained by deleting the profiles in regard to such old topics .
  • the profile deletion program 125 deletes the profiles designated with the user 107 from the user profile storing area (S1601).
  • Fig. 16 is a schematic diagram showing the profile management image in the profile management method of the present invention.
  • a profile monitor result 1702 of the relevant user is displayed in the terminal display image 1701 of the user 107.
  • the information "The profile A is similar to the profile B.” due to the inter-profile similarity monitor program 126 and the information "Information conforming to the profile C is not generated recently.” due to the profile validity monitor program 127 are displayed. Simultaneously, moreover, contents of these profiles are presented as the reference information.
  • the user 107 is therefore capable of determining how optimize the profiles by referring to these profile monitor results 1702 and then requesting such optimization to the system.
  • the user 107 has obtained the information indicating that the profile A and the profile B are similar and therefore thought to form a profile by integrating these profiles A and B.
  • the user 107 depresses the "Integrate" button 1706 by checking the check box 1703 of the profile A and the check box 1704 of the profile B with a pointing device such as a mouse.
  • the profile integration program 123 is driven to set a profile having integrated the profile A and profile B and thereafter the old profiles A and B are deleted.
  • the user 107 determines that the profile C is already unnecessary profile by checking the contents of profile C
  • the user 107 is requested to check only the check box 1705 of the profile C and then depresses the "Delete" button 1708.
  • the profile deletion program 125 is driven and the profile C is then deleted.
  • the user 107 is easily capable of detecting the conditions of the profiles being set. Moreover, the user 107 also can execute, with simplified manipulation, the re-arrangement of profiles such as optimization of profiles and deletion of useless profiles.
  • the information filtering system structured with a display 100, a keyboard 101, a central processing unit (CPU) 102, a main memory 104 and a bus 103 connecting these elements can be located at the area on any network provided at the at the intermediate area between the document information distribution source 106 and the communication line 105 and at the intermediate area between the communication line 105 and user 107 shown in Fig. 1.
  • inter-profile similarity monitor program 126 profile integration program 123 and profile specialization program 124 are provided for the processes of a plurality of profiles set by the user 107 but these programs can also be used for the processes of profiles preset by different users 107.
  • the information filtering system explained in this embodiment includes all of the profile integration program 123, profile specialization program 124 and profile deletion program 125, but it is also possible to realize the information filtering system including any desired combination of these programs.
  • the inter-profile similarity monitor program 126, profile validity monitor program 127, profile integration program 123 and profile specialization program 124 are installed in the information filtering system but this information filtering system can also be utilized for the user to store a plurality of profiles in the document searching system in which the user can search the document database in the desired timing.
  • the present invention can provide a profile management method for obtaining the distribution result without any noise and omission in the information filtering by detecting existence of a plurality of similar profiles and old profiles, then notifying this fact to the user for urging the relevant user to take an adequate measure in such information filtering for filtering and presenting document information using profiles as the search condition data.
  • the present invention can also provide a profile management method to prevent holding of useless profiles by adequately and easily optimizing and deleting old profiles and useless profiles, thereby to permit the user to effectively set the profiles and to maintain the performance of the system by eliminating reference to the useless profiles.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
EP02250562A 2001-06-28 2002-01-28 Verfahren und Programm zur Verwaltung von Profilen zum Filtern von Informationen Withdrawn EP1271357A3 (de)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2001195859 2001-06-28
JP2001195859A JP2003016094A (ja) 2001-06-28 2001-06-28 情報フィルタリングに用いるプロファイル管理方法、および、プロファイル管理プログラム

Publications (2)

Publication Number Publication Date
EP1271357A2 true EP1271357A2 (de) 2003-01-02
EP1271357A3 EP1271357A3 (de) 2005-10-05

Family

ID=19033767

Family Applications (1)

Application Number Title Priority Date Filing Date
EP02250562A Withdrawn EP1271357A3 (de) 2001-06-28 2002-01-28 Verfahren und Programm zur Verwaltung von Profilen zum Filtern von Informationen

Country Status (3)

Country Link
US (1) US20030009440A1 (de)
EP (1) EP1271357A3 (de)
JP (1) JP2003016094A (de)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1837823A1 (de) * 2006-03-22 2007-09-26 Robert Bosch Gmbh Verfahren zur Steuerung einer Anwendung

Families Citing this family (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4186474B2 (ja) * 2002-02-20 2008-11-26 富士通株式会社 プロファイル結合装置、方法及びプログラム
US8412706B2 (en) 2004-09-15 2013-04-02 Within3, Inc. Social network analysis
US8635217B2 (en) * 2004-09-15 2014-01-21 Michael J. Markus Collections of linked databases
US20070260599A1 (en) * 2004-09-15 2007-11-08 Mcguire Heather A Social network analysis
US8577886B2 (en) * 2004-09-15 2013-11-05 Within3, Inc. Collections of linked databases
US8880521B2 (en) * 2004-09-15 2014-11-04 3Degrees Llc Collections of linked databases
JP2006134236A (ja) * 2004-11-09 2006-05-25 Canon Inc プロファイル取得方法、装置、プログラム、および、記憶媒体
US7386545B2 (en) 2005-03-31 2008-06-10 International Business Machines Corporation System and method for disambiguating entities in a web page search
US8631006B1 (en) * 2005-04-14 2014-01-14 Google Inc. System and method for personalized snippet generation
US8453044B2 (en) * 2005-06-29 2013-05-28 Within3, Inc. Collections of linked databases
US8190681B2 (en) 2005-07-27 2012-05-29 Within3, Inc. Collections of linked databases and systems and methods for communicating about updates thereto
US10395326B2 (en) * 2005-11-15 2019-08-27 3Degrees Llc Collections of linked databases
KR100856916B1 (ko) * 2007-01-16 2008-09-05 (주)첫눈 관심사를 반영하여 추출한 정보 제공 방법 및 시스템
JP2008217370A (ja) * 2007-03-02 2008-09-18 Nec Corp プロファイル登録システム、プロファイル登録方法およびプロファイル登録プログラム
US20110035387A1 (en) * 2009-08-10 2011-02-10 Telcordia Technologies, Inc. System and method for the controlled introduction of noise to information filtering
CN106445961B (zh) * 2015-08-10 2021-02-23 北京奇虎科技有限公司 新闻推送方法及装置
JP6785583B2 (ja) * 2016-06-13 2020-11-18 株式会社日立製作所 淡水化装置

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5446891A (en) * 1992-02-26 1995-08-29 International Business Machines Corporation System for adjusting hypertext links with weighed user goals and activities
US6208989B1 (en) * 1994-06-21 2001-03-27 International Business Machines Corporation Facility for the intelligent selection of information objects
US6308175B1 (en) * 1996-04-04 2001-10-23 Lycos, Inc. Integrated collaborative/content-based filter structure employing selectively shared, content-based profile data to evaluate information entities in a massive information network

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5694594A (en) * 1994-11-14 1997-12-02 Chang; Daniel System for linking hypermedia data objects in accordance with associations of source and destination data objects and similarity threshold without using keywords or link-difining terms
US5758257A (en) * 1994-11-29 1998-05-26 Herz; Frederick System and method for scheduling broadcast of and access to video programs and other data using customer profiles
US6029195A (en) * 1994-11-29 2000-02-22 Herz; Frederick S. M. System for customized electronic identification of desirable objects
US6112186A (en) * 1995-06-30 2000-08-29 Microsoft Corporation Distributed system for facilitating exchange of user information and opinion using automated collaborative filtering
US5867799A (en) * 1996-04-04 1999-02-02 Lang; Andrew K. Information system and method for filtering a massive flow of information entities to meet user information classification needs
US6092091A (en) * 1996-09-13 2000-07-18 Kabushiki Kaisha Toshiba Device and method for filtering information, device and method for monitoring updated document information and information storage medium used in same devices
US5778363A (en) * 1996-12-30 1998-07-07 Intel Corporation Method for measuring thresholded relevance of a document to a specified topic
US6003020A (en) * 1997-10-30 1999-12-14 Sapient Health Network Intelligent profiling system
JP4198786B2 (ja) * 1998-06-30 2008-12-17 株式会社東芝 情報フィルタリングシステム、情報フィルタリング装置、映像機器および情報フィルタリング方法
US6757720B1 (en) * 1999-05-19 2004-06-29 Sun Microsystems, Inc. Profile service architecture
JP2001256253A (ja) * 2000-03-13 2001-09-21 Kddi Corp 文書フィルタリング方法および装置
US6463428B1 (en) * 2000-03-29 2002-10-08 Koninklijke Philips Electronics N.V. User interface providing automatic generation and ergonomic presentation of keyword search criteria

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5446891A (en) * 1992-02-26 1995-08-29 International Business Machines Corporation System for adjusting hypertext links with weighed user goals and activities
US6208989B1 (en) * 1994-06-21 2001-03-27 International Business Machines Corporation Facility for the intelligent selection of information objects
US6308175B1 (en) * 1996-04-04 2001-10-23 Lycos, Inc. Integrated collaborative/content-based filter structure employing selectively shared, content-based profile data to evaluate information entities in a massive information network

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
CETINTEMEL U ET AL: "Self-adaptive user profiles for large-scale data delivery" DATA ENGINEERING, 2000. PROCEEDINGS. 16TH INTERNATIONAL CONFERENCE ON SAN DIEGO, CA, USA 29 FEB.-3 MARCH 2000, LOS ALAMITOS, CA, USA,IEEE COMPUT. SOC, US, 29 February 2000 (2000-02-29), pages 622-633, XP010378759 ISBN: 0-7695-0506-6 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1837823A1 (de) * 2006-03-22 2007-09-26 Robert Bosch Gmbh Verfahren zur Steuerung einer Anwendung

Also Published As

Publication number Publication date
EP1271357A3 (de) 2005-10-05
US20030009440A1 (en) 2003-01-09
JP2003016094A (ja) 2003-01-17

Similar Documents

Publication Publication Date Title
EP1271357A2 (de) Verfahren und Programm zur Verwaltung von Profilen zum Filtern von Informationen
KR100304335B1 (ko) 키워드 추출 시스템 및 그를 사용한 문서 검색 시스템
US20080201297A1 (en) Method and System for Determining Relation Between Search Terms in the Internet Search System
US5713016A (en) Process and system for determining relevance
EP1225517A2 (de) System und Methode zum rechnerbasierten Suchen nach relevanten Texten
JP5232260B2 (ja) 話題抽出装置及びプログラム
JP2009545809A (ja) インクリメンタルに更新可能な変形ナイーブベイズクエリ分類器を使用したランク付け関数
JP2004213675A (ja) 構造化ドキュメントの検索
JP4453437B2 (ja) 検索キーワード順位付け方法及び装置及びプログラム
KR20090017268A (ko) 사용자의 관심 키워드를 업데이트 하는 방법과 그 방법을수행하기 위한 시스템
JP4569380B2 (ja) ベクトル生成方法及び装置及びカテゴリ分類方法及び装置及びプログラム及びプログラムを格納したコンピュータ読み取り可能な記録媒体
Bar-Yossef et al. Efficient search engine measurements
US20050027678A1 (en) Computer executable dimension reduction and retrieval engine
CN102890690A (zh) 目标信息搜索方法和装置
JP3350594B2 (ja) 検索装置
JP2018005633A (ja) 関連コンテンツ抽出装置、関連コンテンツ抽出方法及び関連コンテンツ抽出プログラム
KR100682552B1 (ko) 사용자의 상황에 따라 검색 엔진별 가중치를 부여하는시스템, 장치, 방법 및 이를 구현할 수 있는 컴퓨터로 읽을수 있는 기록 매체
JP2010286888A (ja) 情報収集システムおよび情報収集方法ならびにそのプログラム
JP5055202B2 (ja) キーワード比較システム、キーワード比較方法およびキーワード比較プログラム
KR20050109106A (ko) 통합 검색 결과를 효율적으로 제공하는 검색 서비스 제공방법 및 시스템
CN109408789B (zh) 一种笔录模板及其生成方法及笔录模板选择系统
JP5595426B2 (ja) 単語抽出方法及び装置及びプログラム
JP4311062B2 (ja) コンテンツ推薦システム
JP2002215647A (ja) テキストマイニング装置及びそれに用いるテキストマイニング方法並びにそれらに用いるプログラム
JP2000035964A (ja) 関連度算出装置および関連度算出プログラムを記録した記憶媒体並びに情報検索システム

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 20020218

AK Designated contracting states

Kind code of ref document: A2

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE TR

AX Request for extension of the european patent

Free format text: AL;LT;LV;MK;RO;SI

PUAL Search report despatched

Free format text: ORIGINAL CODE: 0009013

AK Designated contracting states

Kind code of ref document: A3

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LI LU MC NL PT SE TR

AX Request for extension of the european patent

Extension state: AL LT LV MK RO SI

AKX Designation fees paid

Designated state(s): GB

REG Reference to a national code

Ref country code: DE

Ref legal event code: 8566

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: THE APPLICATION IS DEEMED TO BE WITHDRAWN

18D Application deemed to be withdrawn

Effective date: 20090127